This is a repository copy of White dwarf-main sequence binaries from LAMOST: the DR5 catalogue.

White Rose Research Online URL for this paper:http://eprints.whiterose.ac.uk/129541/

Version: Accepted Version

Article:

Ren, J.J., Rebassa-Mansergas, A., Parsons, S.G. orcid.org/0000-0002-2695-2654 et al. (4more authors) (2018) White dwarf-main sequence binaries from LAMOST: the DR5 catalogue. Monthly Notices of the Royal Astronomical Society, 477 (4). pp. 4641-4654. ISSN 0035-8711

https://doi.org/10.1093/mnras/sty805

eprints@whiterose.ac.ukhttps://eprints.whiterose.ac.uk/

Reuse

Items deposited in White Rose Research Online are protected by copyright, with all rights reserved unless indicated otherwise. They may be downloaded and/or printed for private study, or other acts as permitted by national copyright laws. The publisher or other rights holders may allow further reproduction and re-use of the full text version. This is indicated by the licence information on the White Rose Research Online record for the item.

Takedown

If you consider content in White Rose Research Online to be in breach of UK law, please notify us by emailing eprints@whiterose.ac.uk including the URL of the record and the reason for the withdrawal request.

WDMS from LAMOST DR5 1

White dwarf-main sequence binaries from LAMOST: the

DR5 catalogue

J.-J. Ren,1⋆ A. Rebassa-Mansergas,2,3 S. G. Parsons,4 X.-W. Liu,5 A.-L. Luo,1

X. Kong,1,6 H.-T. Zhang11Key Laboratory of Optical Astronomy, National Astronomy Observatories, Chinese Academy of Sciences, Beijing 100012, P. R. China2Departament de Fısica, Universitat Politecnica de Catalunya, c/Esteve Terrades 5, E-08860 Castelldefels, Spain3Institute for Space Studies of Catalonia, c/Gran Capita 2–4, Edif. Nexus 201, 08034 Barcelona, Spain4Department of Physics and Astronomy, University of Sheffield, Sheffield S3 7RH, UK5South-Western Institute for Astronomy Research, Yunnan University, Kunming, Yunnan 650091, P. R. China6University of Chinese Academy of Sciences, Beijing 10049, P. R. China

Accepted XXX. Received YYY; in original form ZZZ

ABSTRACTWe present the data release (DR) 5 catalogue of white dwarf-main sequence (WDMS) binaries fromthe Large Area Multi-Object fiber Spectroscopic Telescope (LAMOST). The catalogue contains876 WDMS binaries, of which 757 are additions to our previous LAMOST DR1 sample and 357are systems that have not been published before. We also describe a LAMOST-dedicated surveythat aims at obtaining spectra of photometrically-selected WDMS binaries from the Sloan DigitalSky Survey (SDSS) that are expected to contain cool white dwarfs and/or early type M dwarfcompanions. This is a population under-represented in previous SDSS WDMS binary catalogues. Wedetermine the stellar parameters (white dwarf effective temperatures, surface gravities and masses,and M dwarf spectral types) of the LAMOST DR5 WDMS binaries and make use of the parameterdistributions to analyse the properties of the sample. We find that, despite our efforts, systemscontaining cool white dwarfs remain under-represented. Moreover, we make use of LAMOST DR5and SDSS DR14 (when available) spectra to measure the Na Iλλ 8183.27, 8194.81 absorption doubletand/or Hα emission radial velocities of our systems. This allows identifying 128 binaries displayingsignificant radial velocity variations, 76 of which are new. Finally, we cross-match our cataloguewith the Catalina Surveys and identify 57 systems displaying light curve variations. These include16 eclipsing systems, two of which are new, and nine binaries that are new eclipsing candidates. Wecalculate periodograms from the photometric data and measure (estimate) the orbital periods of 30(15) WDMS binaries.

Key words: binaries: close – binaries: spectroscopic – white dwarfs – stars: low-mass

1 INTRODUCTION

Detached binary stars containing a white dwarf (WD) pri-mary and a main sequence companion star are referred toas WD-main sequence (WDMS) binaries. Depending on theorbital separations of the main sequence binaries from whichthey descend, the previous evolution of WDMS binaries fol-lows two main different paths. Thus, in approximately 75 percent of the cases the orbital separations are wide enough forthe WD precursors to evolve as single stars (de Kool 1992;Willems & Kolb 2004). In the remaining cases the main se-quence binary components are close enough for the systemsto experience dynamically unstable mass transfer interac-tions when the more massive star evolves into a red giant

⋆ E-mail: jjren@nao.cas.cn (JJR).

or asymptotic giant. This generally results in a common en-velope (CE) phase (Iben & Livio 1993; Webbink 2008) inwhich friction of the binary components with the materialof the envelope leads to a dramatic decrease of the binaryseparation. The orbital energy released during this processis used to eventually expel the envelope (Passy et al. 2012;Ricker & Taam 2012). Close WDMS binaries that evolvethrough a CE phase are known as post-CE binaries orPCEBs. Whilst, stricly speaking, a PCEB refers to any typeof binary that evolved through CE evolution – e.g. hot subd-warf B stars with close low-mass main sequence companions(Han et al. 2002, 2003; Heber 2016) – in this work we willonly consider a PCEB as a close WDMS binary.

The orbital period distribution of WDMS binaries isbi-modal. Numerical simulations predict the orbital pe-riods of wide systems that did not evolve through a

Downloaded from https://academic.oup.com/mnras/advance-article-abstract/doi/10.1093/mnras/sty805/4956533by gueston 12 April 2018

2 J.-J. Ren et al.

CE to range between a few hundred to several thou-sand days (Willems & Kolb 2004; Camacho et al. 2014;Cojocaru et al. 2017). This is observationally confirmed byFarihi et al. (2010). The observed PCEB orbital period dis-tribution displays a clear concentration of systems at ∼ 8 h(Miszalski et al. 2009; Nebot Gomez-Moran et al. 2011), al-though PCEBs with periods as long as 10 days have beenalso identified (e.g. Rebassa-Mansergas et al. 2012b).

After the CE phase, PCEBs evolve to even shorterorbital periods through angular momentum loss drivenby magnetic braking and/or gravitational wave emis-sion. Therefore, they may undergo a second phaseof CE evolution, leading to double-degenerate WDs(Rebassa-Mansergas et al. 2017; Breedt et al. 2017;Kilic et al. 2017), or enter a semidetached state andbecome cataclysmic variables (Gansicke et al. 2009;Pala et al. 2017) or super-soft X-ray sources (Parsons et al.2015b). Double degenerate WDs, cataclysmic variablesand super-soft X-ray sources are considered to be possibleprogenitors of type Ia supernova (Langer et al. 2000;Han & Podsiadlowski 2004; Wang et al. 2010; Wang & Han2012), which are of important interest for cosmologicalstudies.

Thanks to the Sloan Digital Sky Survey (SDSS;York et al. 2000; Stoughton et al. 2002) the number ofspectroscopically confirmed WDMS binaries has dramat-ically increased during the last decade (Silvestri et al.2006; Rebassa-Mansergas et al. 2007; Heller et al.2009; Rebassa-Mansergas et al. 2010, 2012a; Liu et al.2012; Rebassa-Mansergas et al. 2013b; Li et al. 2014;Rebassa-Mansergas et al. 2016a). Thus, the currentmost updated SDSS WDMS catalogue includes 3 294binaries1, which is by far the largest and most homo-geneous sample of compact binaries in the literature(Rebassa-Mansergas et al. 2016a). Among the 3 294 SDSSWDMS, observational follow-up studies have resulted in theidentification of ∼1000 wide binaries that did not interactand more than 200 PCEBs (Rebassa-Mansergas et al.2007; Schreiber et al. 2008, 2010; Rebassa-Mansergas et al.2011, 2016a), of which 90 have available orbital periods(Rebassa-Mansergas et al. 2008; Nebot Gomez-Moran et al.2011; Rebassa-Mansergas et al. 2012b) and 71 are eclipsing(Nebot Gomez-Moran et al. 2009; Pyrzas et al. 2009, 2012;Parsons et al. 2013a, 2015a).

The superb sample of SDSS WDMS binaries hasled to many and diverse advances in the field ofastrophysics. Among these we emphasise the follow-ing: providing constraints on theories of CE evolution(Davis et al. 2010; Zorotovic et al. 2010; De Marco et al.2011), analysing the Galactic properties of WDMS bi-naries through population synthesis studies aimed atreproducing the ensemble properties of the observedpopulation (Toonen & Nelemans 2013; Camacho et al.2014; Zorotovic et al. 2014; Cojocaru et al. 2017), testingthe age-metallicity relation in the solar neighbourhood(Rebassa-Mansergas et al. 2016b), identifying pulsating low-mass WDs in detached WDMS binaries (Pyrzas et al.2015), providing evidence for disrupted magnetic braking(Schreiber et al. 2010; Zorotovic et al. 2016), studying the

1 https://sdss-wdms.org/

origin of low-mass WDs (Rebassa-Mansergas et al. 2011),investigating the age-rotation-activity relation of low-massM stars (Rebassa-Mansergas et al. 2013a; Skinner et al.2017), constraining the critical binary star separation fora planetary system origin of WD pollution (Veras et al.2018), testing the mass-radius relation of low-mass mainsequence stars and WDs (Parsons et al. 2012a,b, 2017)and investigating the origin of circumbinary giant planetsaround PCEBs (Zorotovic & Schreiber 2013; Marsh et al.2014; Parsons et al. 2014; Schleicher et al. 2015; Bours et al.2016).

However, despite the importance of the SDSS WDMSbinary sample, it is important to emphasise that it suf-fers from serious selection effects (Rebassa-Mansergas et al.2016a). This is because the SDSS focused mainly on spec-troscopic observations of quasars and galaxies, which over-lap in colour space with hot WDs (Richards et al. 2002;Smolcic et al. 2004). Hence, WDMS containing cool WDsand early-type companions are under-represented. In or-der to overcome this issue, Rebassa-Mansergas et al. (2012a)presented a dedicated survey within SEGUE (The SloanExtension for Galactic Understanding and Exploration;Yanny et al. 2009) that obtained spectra for 291 WDMSbinaries containing cool WDs and/or early type compan-ions. Moreover, Rebassa-Mansergas et al. (2013b) photo-metrically selected 3419 WDMS candidates for containingcool WDs.

The Large sky Area Multi-Object fiber SpectroscopicTelescope (LAMOST) survey is a large-scale spectroscopicsurvey which follows completely different target selection al-gorithms than SDSS, hence opening the possibility for iden-tifying WDMS binaries that help in overcoming the selectioneffects of the SDSS sample. Ren et al. (2013, 2014) identi-fied a total of 121 LAMOST WDMS binaries from the datarelease (DR) 1 of LAMOST (Luo et al. 2012). They foundthat the intrinsic properties of the LAMOST WDMS binarysample were different from those of the SDSS population.In particular, LAMOST WDMS binaries are found at con-siderably shorter distances and are dominated by systemscontaining early-type companions and hot WDs (Ren et al.2014). LAMOST WDMS thus represent an important addi-tion to the current known spectroscopic catalogue of SDSSWDMS binaries. However, the number of WDMS containingcool WDs remains still under-represented, and the numberof LAMOSTWDMS binaries identified so far is rather small.

Here we present an updated LAMOST WDMS binarycatalogue based on the most recent DR of LAMOST –DR5, to thus increase the current small number of LAM-OST WDMS binaries. Moreover, we present a dedicatedLAMOST survey for obtaining spectra of as many as pos-sible photometrically selected SDSS WDMS binaries fromRebassa-Mansergas et al. (2013b) that presumably containcool WDs. We measure the stellar parameters from theLAMOST spectra of all our identified objects and perform astatistical analysis of their intrinsic properties. We also de-termine the radial velocities (RVs) of each binary with theaim of detecting PCEBs. Finally, we cross-match our sam-ple with photometric data from the Catalina Sky Survey(CSS) and Catalina Real Time Transient Survey (CRTS;Drake et al. 2009) in order to search for WDMS binariesdisplaying light curve variations (e.g. eclipsing systems).

Downloaded from https://academic.oup.com/mnras/advance-article-abstract/doi/10.1093/mnras/sty805/4956533by gueston 12 April 2018

WDMS from LAMOST DR5 3

2 THE DR5 OF LAMOST

LAMOST is a quasi-meridian reflecting Schmidt telescope of∼4 meter effective aperture and a field of view of 5 degrees indiameter (Wang et al. 1996; Su et al. 2004; Cui et al. 2012).LAMOST is located in Xinglong station of National Astro-nomical Observatories of Chinese Academy of Sciences. Be-ing a dedicated large-scale survey telescope, LAMOST uses4 000 fibres to obtain spectra of celestial objects as well assky background and calibration sources in one single ex-posure. To that end, LAMOST uses 16 fiber-fed spectro-graphs each accommodating 250 fibers. Each spectrographis equipped with two CCD cameras of blue and red chan-nels that simultaneously provide blue and red spectra of the4 000 selected targets, respectively. The LAMOST spectracover the entire optical wavelength range (≃ 3700 – 9000A)at a resolving power R∼ 1800.

From 2012 September, LAMOST has been carrying outa five-year Regular Survey. Before that, there was a one-yearPilot Survey preceded by a two-year commissioning phase.The LAMOST Regular Survey consists of two components(Zhao et al. 2012): the LAMOST Extra-Galactic Survey ofgalaxies (LEGAS) that aims at studying the large scalestructure of the universe, and the LAMOST Experimentfor Galactic Understanding and Exploration (LEGUE) thataims at obtaining millions of stellar spectra in order to studythe structure and evolution of the Milky Way (Deng et al.2012). LEGUE is sub-divided into three sub-surveys: thespheroid, the anticentre (Liu et al. 2014; Yuan et al. 2015;Xiang et al. 2017) and the disc surveys. The five-year Reg-ular Survey finished in June 16th 2017 and the spectra havebeen released internally to the Chinese scientific commu-nity through the DR5 of LAMOST2. The raw spectra areprocessed with the LAMOST two-dimensional (2D) pipeline(Luo et al. 2012, 2015), which includes dark and bias sub-tractions, cosmic ray removal, one-dimensional (1D) spec-tral extraction, merging sub-exposures (note that each plateis generally observed consecutively for a minimum of threetimes), and finally, splicing the sub-spectra from the blue-and red channels of the spectrographs, respectively. TheLAMOST 1D pipeline is then carried out to perform spec-tral classification and to measure the redshift (or radial ve-locity) of each spectrum. There are 4 119 plates observed inLAMOST DR5, thus including a total of 8 952 297 spectra,of which 7 930 178 are stars, 152 608 are galaxies, 50 132 arequasars, and 819 379 are classified as unknowns.

3 THE LAMOST DR5 WDMS BINARYCATALOGUE

In this section we describe how we identify WDMS binarieswithin LAMOST DR5 and we estimate the completeness ofthe sample.

3.1 Identification of WDMS binaries

In Ren et al. (2014) we developed a fast and efficient rou-tine based on the wavelet transform (WT; Chui 1992) toidentify WDMS binaries from LAMOST DR1. The analysis

2 http://dr5.lamost.org/

4000 4500 5000

Wavelength (Å)

1000

1500

Flux

0 4 8 12 16Data Number

34

56

cA5

7000 8000 9000

Wavelength (Å)

1000

1500

Flux

0 4 8 12Data Number

812

cA7

Figure 1. A WT decomposition example of a LAMOST DA/Mbinary (J161713.52+215517.5). The top panels show the blue(left) and red (right) spectrum of the binary. The vertical grey

dashed lines indicate the spectral regions selected to perform thewavelet transform (the 3910–4422A range in the blue and the6800–8496A range in the red). The bottom panels display theapproximation coefficients (CA) as a function of data number ofpoints after applying the WT (5th iteration in blue, left panel;7th iteration in red, right panel).

unit of the WT is the local flux of the spectrum, i.e. the se-lected spectral features. In other words, the WT recognisesthe spectral features rather than the global continuum ofa given spectrum. This is done by decomposing a consid-ered spectral range (covering the desired spectral features)into approximation signals, often referred to as approxima-tion coefficients. The wavelength values under the consid-ered spectral region are converted into data points. Thisdecomposition process can be iterated by decomposing theapproximation coefficients into successive approximation co-efficients (thus reducing by half the number of data pointsin each iteration) until the decomposition level is satisfac-tory. This occurs when the spectral features of a WDMSbinary spectrum can be identified in the approximation co-efficients as compared to a non-WDMS binary spectrum inwhich the approximation coefficients are dominated by con-tinuum emission and/or by spectral features different fromthose of WDMS binaries. The outcome of a WT can be henceconsidered as a smoothed version of the spectrum (for com-parative purposes). Because of its high efficiency in identi-fying spectral features, the WT is very suitable to identifyWDMS binaries among low signal-to-noise (SN) ratio spec-tra.

WDMS binaries containing a DA (hydrogen-rich) WDand a low-mass main sequence star have two obvious spec-tral features: the Balmer lines in the blue band arising fromthe WD and the molecular absorption bands in the redthat are typical of M dwarfs. We hence applied the WTto the following spectral regions: the 3910–4422A rangein the blue band that covers the Hβ to Hǫ Balmer linesand the 6800–8496A range which covers a large numberof TiO and VO molecular bands. By choosing these twospectral features we focus our search on the identificationof WDMS binaries containing a DA WD and a M dwarfcompanion (hereafter DA/M binaries; see an example ofWT applied to a DA/M binary in Figure 1). The mainreason for this choice is that DA/M are by far the most

Downloaded from https://academic.oup.com/mnras/advance-article-abstract/doi/10.1093/mnras/sty805/4956533by gueston 12 April 2018

4 J.-J. Ren et al.

Table 1. The catalogue of LAMOST DR5 WDMS binaries, including a total of 1 150 spectra of 876 unique targets. Here we list thecoordinates, SDSS ugriz magnitudes (when available), LAMOST spectral identifiers (plate, spectrograph, fibre IDs and modified Juliandate MJD), signal-to-noise ratio in five different bands (SNu, SNg, SNr , SNi, SNz) and stellar parameters (the WD effective temperature,surface gravity and mass, and the spectral type of the M dwarf companion). We also provide Simbad classifications (when available) andindicate if the target is identified by the WT (fWT=1, yes; 0, no), if it is a photometrically selected SDSS WDMS binary that obtained aLAMOST spectrum (fPhot=1, yes; 0, no), if it is part of the SDSS DR12 WDMS spectroscopic catalogue (fSDSS=1, yes; 0, no), if it is aWDMS binary candidate (fCAND=1, candidates; 0, genuine DA/M) and/or if it is part of the LAMOST DR1 WDMS binary catalogue(fDR1=0, not in DR1; 10, in DR1 genuine DA/M sample, not in DR1 DA/M candidate sample; 11, in DR1 DA/M candidate sample).The entire table is provided in the electronic version of the paper. We use ‘−’ to indicate that no magnitude or stellar parameter isavailable.

Jname fWT fPhot fSDSS fCAND fDR1 RA Dec u Err g Err r Err i Err z Err() () (mag) (mag) (mag) (mag) (mag) (mag) (mag) (mag) (mag) (mag)

J000448.23+343627.4 1 0 0 0 00 1.2009999 34.6076320 − − − − − − − − − −

J000612.32+340358.3 1 1 0 0 00 1.5513397 34.0661980 18.954 0.029 18.327 0.008 17.794 0.008 16.535 0.005 15.713 0.007

J000729.32+023124.5 1 0 0 0 00 1.8721920 2.5234990 19.912 0.042 19.092 0.010 18.379 0.008 17.494 0.007 16.937 0.011J001258.26+062617.9 1 0 0 0 00 3.2427542 6.4383056 18.505 0.016 17.987 0.006 18.001 0.007 17.528 0.007 17.088 0.011J001733.59+004030.4 1 0 1 0 00 4.3899705 0.6751216 20.200 0.051 19.517 0.014 18.982 0.013 17.925 0.009 17.166 0.014J001752.63+332424.9 1 1 0 0 00 4.4693105 33.4069250 18.582 0.015 17.625 0.005 16.519 0.004 15.616 0.004 15.069 0.005J001752.63+332424.9 1 1 0 0 00 4.4693105 33.4069250 18.582 0.015 17.625 0.005 16.519 0.004 15.616 0.004 15.069 0.005J002237.90+334322.1 1 0 0 0 10 5.6579420 33.7228090 19.548 0.033 17.361 0.005 15.985 0.003 14.590 0.004 13.858 0.004J002407.31+054856.4 1 1 0 0 00 6.0304700 5.8156830 17.068 0.008 16.393 0.004 15.851 0.004 14.808 0.004 14.128 0.004J002633.13+390904.0 1 1 0 0 00 6.6380824 39.1511210 16.348 0.006 15.905 0.003 16.011 0.004 15.831 0.004 15.327 0.005J002633.14+390904.0 1 1 0 0 00 6.6385030 39.1508170 16.348 0.006 15.905 0.003 16.011 0.004 15.831 0.004 15.327 0.005

MJD Plate spID fiberID SNu SNg SNr SNi SNz TeffWD Err log gWD Err MWD Err Sp Simbad(K) (K) (dex) (dex) (M⊙) (M⊙)

56948 M31001N35M2 05 156 8.66 33.80 30.05 49.00 32.49 16717 245 8.02 0.06 0.63 0.03 4 −

56948 M31001N35M2 01 139 3.92 12.66 22.09 47.00 50.04 28389 1448 7.87 0.23 0.58 0.11 4 −

55893 F9302 01 235 1.72 2.52 4.11 7.63 6.58 − − − − − − 1 −

56602 EG001605N080655M01 02 066 1.52 3.01 2.25 3.48 3.23 − − − − − − 4 −

56978 EG001639N015102M01 05 141 1.52 5.07 7.31 16.86 14.32 − − − − − − 3 DA+dM56952 M31005N35M1 02 120 2.44 13.88 35.84 71.58 60.63 − − − − − − 1 −

57717 M31007N33B2 10 109 3.66 16.36 34.31 71.88 63.27 29727 1312 8.34 0.25 0.82 0.14 2 −

56255 VB007N33V1 03 085 2.23 3.49 2.78 1.72 3.15 − − − − − − − DA+M56621 EG002616N034932B01 11 020 8.25 27.85 40.10 86.82 82.41 14899 395 8.41 0.08 0.85 0.05 3 −

57280 M31007N36M1 11 163 1.80 3.81 4.02 4.78 4.40 − − − − − − 9 DA4.656911 HD002951N381926B01 15 227 4.28 9.16 5.54 4.01 3.84 − − − − − − − DA4.6

Figure 2. Example spectra of LAMOST DA/M binaries identi-fied by the WT in descendant order of SN ratio (in the r band).

common among WDMS binaries, including ∼ 80 per cent ofall systems (Rebassa-Mansergas et al. 2016a). This is be-cause of the following two reasons. First, DA WDs comprise∼ 80 per cent of all single WDs (Kepler et al. 2015), henceit is not surprising that they are the most common WDsin binaries too. Second, because of selection effects, mainsequence stars of spectral type earlier than M generally out-shine the WDs in the optical, which implies only a handfulof WDs with late K-type companions have been identified(Rebassa-Mansergas et al. 2016a). We will pursue the iden-tification of LAMOST WDMS binaries harboring other WDtypes (DB, DQ, DZ, etc) as well as late K-type companionselsewhere. Of course, in order for our WT to efficiently iden-tify DA/M binaries we require the two components (DAWDplus M dwarf) to be visible in the spectrum. This impliesthat systems in which one of the two components clearlydominates the spectral energy distribution will be harder(or even impossible) to detect. We note however that this is-sue makes the detection of such systems to be very difficultby any method (see for example Rebassa-Mansergas et al.2010).

After applying the WT method to the ∼9 million LAM-OST DR5 spectra, we obtained an initial list of 29 269WDMS binary candidate spectra that we visually inspected.This resulted in 776 spectra (579 unique systems) that weconsidered as genuine DA/M binaries, and 47 spectra (46unique systems) that we catalogued as DA/M candidates.The majority of spectra selected by the WT that we did notconsider DA/M binaries corresponded to single MS stars andto the superposition in the line of sight of two main sequence

Downloaded from https://academic.oup.com/mnras/advance-article-abstract/doi/10.1093/mnras/sty805/4956533by gueston 12 April 2018

WDMS from LAMOST DR5 5

stars. Table 1 lists all the genuine and candidate DA/M bi-naries identified by the WT, where we also include LAM-OST spectroscopic identifiers and modified Julian dates ofthe observations. In Figure 2 we show four example spectraof LAMOST DA/M binaries of different SN ratio.

3.2 Completeness of the LAMOST DR5 DA/Mbinary sample

In Ren et al. (2014) we showed the WT method is highly ef-ficient (∼90 per cent) in identifying DA/M binaries and thatthe spectra it fails to identify are either clearly dominatedby one of the stellar components and/or are associated tovery low SN ratio (< 5). In order to estimate the complete-ness of the LAMOST DR5 DA/M binary catalogue (i.e. thenumber of DA/M binaries observed by LAMOST that theWT successfully identified), we cross-matched the entire ∼9million spectroscopic data base of LAMOST with the SDSSDR12 spectroscopic catalogue of WDMS binaries. We found447 objects in common (583 LAMOST spectra).

111 (153 spectra) of the 447 objects (583 spectra) werenon-DA/M binaries that the WT failed to identify for obvi-ous reasons: the spectral features of these WDMS binarieswere different from those of DA/M binaries. In order to com-pile a catalogue as complete as possible, we added the 111objects (105 genuine WDMS, 143 spectra; 6 WDMS binarycandidates, 7 spectra) to our list.

The remaining 330 systems were DA/M binaries (424spectra), 233 (312 spectra) of which were identified by theWT and 97 (112 spectra) were missed. This translated into acompleteness of ∼70 per cent, a value which is considerablylower than the ∼90 per cent claimed by Ren et al. (2014).We repeated the exercise excluding all spectra of SN ratiobelow 5 and found the number of DA/M binaries identifiedand missed by the WT were 156 (188 spectra) and 37 (42spectra), respectively, i.e. a completeness of ∼80 per cent.We visually inspected the 42 missed spectra and realisedthat 19 of them were clearly dominated by the flux contri-bution of the WD and 15 by the flux contribution of theM dwarf. We also found 5 broken spectra. Excluding thebroken spectra and those dominated by one of the stellarcomponents, the completeness increased to ∼98 per cent.We added the 97 objects (93 genuine WDMS, 108 spectra;4 candidate WDMS, 4 spectra) the WT missed to our listof LAMOST DR5 WDMS binaries.

We conclude the WT is highly efficient at identifyingDA/M binaries and that the LAMOST DR5 DA/M binarycatalogue is ∼98 per cent complete. However, the complete-ness drops to ∼70 per cent when considering spectra clearlydominated by the flux contribution of one of the stellar com-ponents and/or spectra of very low SN ratio (<5). Unfortu-nately, identifying DA/M binary spectra of such characteris-tics is challenging for any method (Rebassa-Mansergas et al.2010).

4 A DEDICATED LAMOST SURVEY OFPHOTOMETRICALLY SELECTED SDSSWDMS BINARIES

The spectroscopic catalogue of SDSS WDMS binaries isseverely biased against systems containing cool WDs and/or

Figure 3. SDSS gri magnitude distributions of the 3 419 photo-metrically selected WDMS binaries of Rebassa-Mansergas et al.(2013b).

early type companions. This is mainly because SDSS ded-icated most of its efforts to obtain spectra of quasars andgalaxies, which overlap in colour space with WDMS binariescontaining hot (&10 000–15 000 K) WDs and/or late type(>M3–4) companions. In order to overcome this selectioneffect, Rebassa-Mansergas et al. (2013b) developed a colourselection criteria that combines optical (SDSS ugriz) plusinfrared (UKIRT Infrared Sky Survey yjhk, Two MicronAll Sky Survey JHK and/or Wide-Field Infrared SurveyExplorer w1w2) magnitudes to photometrically select 3 419WDMS binary candidates for harboring cool WDs and/ordominant (M dwarf) companions in their spectra. Visual in-spection of 567 photometric candidates with available SDSSspectra allowed Rebassa-Mansergas et al. (2013b) to con-clude that 84 per cent of entire the photometric sample areexpected to be genuine WDMS binaries, and that 71 per centshould contain WDs of effective temperatures below 10 000–15 000K and M dwarf companions of spectral types concen-trated at ∼M2–3.

In this section we present a dedicated LAMOST sur-vey that so far has obtained spectra of 622 photometricallyselected SDSS WDMS binaries.

4.1 The LAMOST observations

Since 2014, the LAMOST target selection algorithm has in-corporated the photometrically selected WDMS binary listof Rebassa-Mansergas et al. (2013b) with the aim of obtain-ing as many WDMS binary spectra as possible. As men-tioned in Luo et al. (2015), the LAMOST observations arecarried out using four different modes, i.e. the selected ob-jects are observed using different plates according to theirmagnitudes: very bright plates (VB, r 6 14mag), brightplates (B, 14mag < r < 16.8mag), medium-bright plates(M, 16.8mag 6 r 6 17.8mag) and faint plates (F, r >

17.8mag). LAMOST DR5 has made use of ∼ 4 000 plates,of which 80 per cent are VB/B plates (46 per cent are VB

Downloaded from https://academic.oup.com/mnras/advance-article-abstract/doi/10.1093/mnras/sty805/4956533by gueston 12 April 2018

6 J.-J. Ren et al.

plates), i.e. most of the observed targets have r magni-tudes below 16.8. Fig 3 shows the gri magnitude distribu-tion of the 3 419 photometrically selected WDMS binariesof Rebassa-Mansergas et al. (2013b), where one can clearlysee the number of systems with r < 16.8 mag is low (∼17per cent). Moreover, it is important to emphasize that, onoccasions, the coordinates of the LAMOST plates and theWDMS binaries do not overlap (see Fig 4). Overall, this im-plies that only a small fraction of photometrically selectedWDMS binaries has been so far observed by LAMOST.

4.2 The success rate of the SDSS photometricsample

The LAMOST observations above described resulted in atotal number of 872 spectra of 622 unique sources fromthe Rebassa-Mansergas et al. (2013b)’s list. We visually in-spected the spectra and identified 298 genuine WDMS bi-naries (436 spectra) and 41 (57 spectra) that we consideredas WDMS binary candidates. The remaining sources werecataclysmic variables (7; 13 spectra), single M dwarfs (161;214 spectra), quasars (23; 36 spectra), main sequence stars(16; 26 spectra), single WDs (4; 4 spectra), main sequenceplus main sequence superpositions (11; 20 spectra), and un-known objects due to the bad quality of their spectra (61;66 spectra).

If we exclude the 61 unknown sources we find that60 per cent of the photometric candidates are indeed gen-uine WDMS binaries or WDMS binary candidates. Thisvalue is low compared to the 84 per cent of successrate claimed by Rebassa-Mansergas et al. (2013b) for thesame sample. However, it is important to emphasise thatRebassa-Mansergas et al. (2013b) showed the success ratevaries from 40 per cent up to 100 per cent depending oncolour space and that the 84 per cent value represents anaverage success rate over the entire colour space. This canclearly be seen in Figure 5, where we illustrate the successrate in form of a density map in the g − r vs. r − i colourspace. In this figure we also show the location of the gen-uine WDMS binaries observed by LAMOST DR5 (greensolid dots) and those that we consider contaminants (i.e.non WDMS binaries). It becomes obvious that the contami-nants are either located in areas of expected low success rateor concentrated at g − r < 1. At such red colours the WDcontribution in the optical spectrum is expected to be ratherlow, which makes it difficult to judge simply by eye whetheror not these objects are real WDMS binaries or single MSstars. Our visual inspection decided for the latter, althoughthe former cannot be ruled out.

We checked how many of the 298 genuine WDMS bi-naries identified in this section were also found by the WTand realised that only 30 objects (a total of 49 spectra) weremissing. We made use of this result to derive (in an alter-native way) the completeness of the LAMOST DR5 WDMSbinary catalogue, which results in this case in ∼ 90 per cent.This value is slightly lower than the 98 per cent we obtainedbefore (see Section 3.2), however it confirms the WT is anefficient tool to identify WDMS binaries among large spec-troscopic samples. We also found that 37 (52 spectra) of the41 (57 spectra) WDMS binary candidates were not identifiedby the WT. Visual inspection revealed most of these candi-date spectra were dominated by the flux of the M dwarf,

some were broken, and others were of low SN ratio. Hence,it is not surprising that the WT was not successful at iden-tifying these spectra.

Among the 30 objects the WT missed (49 spectra), 14(25 spectra) were found cross-matching the entire LAMOSTdata base with the SDSSDR12WDMS binary catalogue (seeSection 3.2). Hence, we added to our list the remaining 16WDMS binaries (24 spectra). In the same way, among the 37WDMS binary candidates (52 spectra) the WT missed, 10(11 spectra) were identified in Section 3.2. Thus, we addedthe remaining 27 (41 spectra) WDMS binary candidates toour LAMOST DR5 catalogue.

5 THE FINAL CATALOGUE OF LAMOST DR5WDMS BINARIES

The WT identified 579 genuine (776 spectra) and 46 can-didate (47 spectra) WDMS binaries (Section 3.1). To thislist we added 97 (112 spectra) DA/M binaries (93 gen-uine binaries, 108 spectra; 4 candidates, 4 spectra) and 111(150 spectra) non DA/M binaries (105 genuine non DA/M,143 spectra; 6 non DA/M candidates, 7 spectra) by cross-matching the SDSS DR12 WDMS catalogue with the entireLAMOST spectroscopic data base (Section 3.2). Finally, weadded 16 (24 spectra) genuine and 27 (41 spectra) candi-date DA/M binaries by visually inspecting the LAMOSTspectra of 622 SDSS photometrically selected targets (Sec-tion 4). This brings the total number of genuine LAMOSTDR5 WDMS binaries to 793 (1 051 spectra) and 83 WDMSbinary candidates (99 spectra), i.e. 876 objects and 1150spectra. The entire catalogue is included in Table 1.

We compared our LAMOST DR5 catalogue to the onewe presented in Ren et al. (2014) and found that 119 objects(194 spectra) were already identified in the DR1 of LAM-OST. Hence, the DR5 catalogue contains 757 (956 spectra)new entries. Moreover, we found that 400 (509 spectra) ob-jects were already discovered in the SDSS DR12 catalogueof Rebassa-Mansergas et al. (2016a), which implies 357 (447spectra) are new systems that have not been published be-fore.

6 CHARACTERIZATION OF THE LAMOSTDR5 WDMS BINARY CATALOGUE

In this section we measure the WD stellar parameters anddetermine the M dwarf spectral subclass of all DA/M bina-ries that are part of our new LAMOST DR5 WDMS binarycatalogue. We also describe the properties of the new DR5sample by analysing the parameter distributions.

6.1 Stellar parameter determinations

We determined the stellar parameters followingthe decomposition/fitting routine described byRebassa-Mansergas et al. (2007). The first step was tofit the DA/M binary spectrum with a two-componentmodel using a set of observed M dwarf and WD templates,which allowed obtaining the spectral type of the M dwarf.Then, the best-fit M dwarf template was subtracted fromthe DA/M binary spectrum and the Balmer lines (Hβ to

Downloaded from https://academic.oup.com/mnras/advance-article-abstract/doi/10.1093/mnras/sty805/4956533by gueston 12 April 2018

WDMS from LAMOST DR5 7

Figure 4. Galactic coordinates of the LAMOST DR5 plates (gray open circles) and the photometrically selected SDSS WDMS binariesof Rebassa-Mansergas et al. (2013b) (solid coloured dots; each colour is associated to a given r magnitude according the colour bar

displayed in the bottom of the figure). The coloured dots with black edges illustrate those WDMS binaries with r < 16.8mag.

Figure 5. The estimated success rate for selecting genuineWDMS binaries based on the colour selection criteria developedby Rebassa-Mansergas et al. (2013b) (black regions represent 100per cent of success rate, white regions 0 per cent). The 873 spec-tra observed by LAMOST DR5 are indicated as solid dots (greenfor genuine WDMS binaries, red for contaminants).

Hǫ) of the residual WD spectrum were fitted using a grid ofDA WD atmosphere models (Koester 2010) to derive theWD effective temperature (Teff) and surface gravity (log g).Given that the equivalent widths of the Balmer lines reacha maximum near Teff = 13 000K (with the exact valuebeing a function of log g), the Teff and log g determinedfrom Balmer line profile fits are subject to a degeneracy.That is, fits of similar quality can be achieved on either side

Figure 6. Comparison between the WD parameters (effectivetemperature, surface gravity and mass) and M dwarf spectralsubclass derived from SDSS and LAMOST spectra of commonDA/M binaries.

of the temperature at which the maximum equivalent widthis occurring. We broke this degeneracy fitting the entireWD residual spectrum (continuum plus Balmer lines).

It has been shown that one-dimensional WD model at-mosphere spectra such as those used in this work yield over-estimated log g values for WDs of effective temperatures be-low ∼12 000K (Koester et al. 2009; Tremblay et al. 2011).Therefore, we applied the corrections of Tremblay et al.(2013), which are based on three-dimensional WD model at-

Downloaded from https://academic.oup.com/mnras/advance-article-abstract/doi/10.1093/mnras/sty805/4956533by gueston 12 April 2018

8 J.-J. Ren et al.

Figure 7. Parameter distributions (WD mass, top panel; WDeffective temperature, top-middle panel; WD surface gravity,bottom-middle panel; secondary M dwarf spectral subtype, bot-tom panel) for the entire LAMOST DR5 DA/M binary catalogue(gray) and for the sub-sample of objects that were photometri-cally selected from the SDSS and observed by LAMOST (black).

mosphere spectra, to our WD parameter determinations. Wethen interpolated the Teff and log g values in the cooling se-quences of Renedo et al. (2010) to derive the WD masses. Incases where more than one spectrum per target were avail-able we averaged the stellar parameters (namely the Teff ,log g and mass of the WD and the spectral type of the Mdwarf). The stellar parameters are provided in Table 1 foreach individual spectrum.

Given that a large number of LAMOST DA/M bina-ries have also been observed by SDSS, we decided to com-pare the stellar parameters determined here with those ob-tained fitting the SDSS spectra. We note that we applied thesame fitting routine to both the LAMOST and SDSS spec-tra. The result can be seen in Figure 6, where we displayonly WDs with relative error in their effective temperaturesbelow 10 per cent (128 objects), and absolute errors below0.15 dex (in log g; 30 objects) and 0.1 M⊙ (in mass; 50 ob-jects). Moreover, we only took into account DA/M binarieswith available M dwarf spectral subclasses (318 objects).Visual inspection reveals the WD parameters are broadlyconsistent with each other. Quantitatively, if we define τ

(Gentile Fusillo et al. 2015):

τ =SDSS value− LAMOST value

√SDSS error2 + LAMOST error2

(1)

then we find that the WD parameters show a comparable

Table 2. KS and χ2 probabilities. Sample 1 refers to the entireLAMOST DR5 catalogue excluding the SDSS photometrically se-lected objects observed by LAMOST, Sample 2 refers to the SDSSphotometrically selected sample observed by LAMOST, Sample3 refers to the LAMOST DR1 catalogue of Ren et al. (2014) andSample 4 refers to the full LAMOST DR5 catalogue.

Sample 1 vs. Sample 2 Sample 3 vs. Sample 4

WD Teff 0.09 0.35WD log g 0.66 0.46WD mass 0.40 0.09M dwarf Sp 0.93 0.56

agreement (τ < 2) in 70 (Teff), 77 (log g) and 82 (mass)per cent of the cases. We can conclude then that the WDparameters derived using LAMOST and SDSS spectra areindeed broadly in agreement.

Visual comparison of the M dwarf spectral subclassesobtained fitting the SDSS and LAMOST spectra (bottomleft panel of Figure 6) seems to indicate a large spread ofvalues for a given spectral subtype. However, only in 12 ofthe 318 cases considered the difference in spectral subtypesis larger than 2, and in 49 cases larger than 1. We visually in-spected the LAMOST spectra of these 49 objects and foundthat they were either dominated by the WD contributionand/or of low SN ratio. These are clear cases in which weexpect higher uncertainties in the determination of the Mdwarf spectral subclasses. In order to quantitatively studythe apparent disagreement further, we obtained the aver-age LAMOST spectral subclass and its standard deviationfixing the SDSS spectral subclasses. The result is shown asred solid dots and red solid lines in the bottom left panelof Figure 6. It becomes clear that, in the majority of cases,the spectral subtypes agree well with each other and thatthe apparent scatter at a given spectral subtype is causedby isolated cases.

6.2 Properties of the LAMOST DA/M binarycatalogue

In the previous section we determined the stellar param-eters of our LAMOST DR5 DA/M binaries. This allowsus to present here the corresponding stellar parameter dis-tributions, which we show in Figure 7 in the following or-der: WD mass (top panel), WD effective temperature (top-middle panel), WD surface gravity (bottom-middle panel)and M dwarf spectral subtype (bottom panel). FollowingSection 6.1, and in order to exclude values subject to largeuncertainties, we exclude WDs with relative error in theireffective temperatures above 10 per cent (we are left with277 objects in the top-middle panel of Figure 7), and ab-solute errors above 0.15 dex in log g (76 objects; middle-bottom panel) and 0.1 M⊙ in mass (116 objects; top panel).In the same way, we only consider DA/M binaries withwell-determined M dwarf spectral subclasses (686 objects;bottom panel of Figure 7). Visual inspection of the param-eter distributions reveals that the WDs are concentrated at∼ 0.6M⊙ (which is equivalent to say they are concentratedat log g ∼ 8 dex), with a steep decline towards higher andlower masses (being the latter WDs that are presumablypart of close binaries, see Section 7 for a further discussion)and with typical effective temperature values between 10 000

Downloaded from https://academic.oup.com/mnras/advance-article-abstract/doi/10.1093/mnras/sty805/4956533by gueston 12 April 2018

WDMS from LAMOST DR5 9

Figure 8. Cumulative parameter distributions of the SDSS photometrically selected sample observed by LAMOST (black dotted lines;left and middle-left panels), the complete DR5 LAMOST sample excluding the SDSS photometrically-selected ones (black solid lines;left and middle-left panels), the LAMOST DR1 catalogue of Ren et al. (2014) (gray dotted lines; right and middle-right panels) and thecomplete LAMOST DR5 catalogue (solid gray lines; right and middle-right panels).

and 40 000K (with a peak at ∼15 000K). The M dwarf spec-tral types are concentrated between M1–M4, with a declinetowards earlier and later types. For completeness, we alsoshow in Figure 7 the parameter distributions that result fromconsidering only the photometrically-selected SDSS DA/Mbinary sample that has been observed by LAMOST. Wenote that we did not include any DA/M binary consideredas candidate in this exercise.

We run Kolmogorov-Smirnov (KS) tests to the WD cu-mulative parameter distributions and apply a χ2 test to thecumulative spectral type distributions resulting from theSDSS photometrically-selected sample and the total sam-ple of LAMOST DR5 DA/M binaries (excluding the SDSSphotometrically selected ones). The cumulative distributionsare shown in the left panels of Figure 8 and the results aregiven in Table 2. The probabilities obtained do not seem toindicate the two samples are statistically different, which iswhat we should expect given that the photometric sampleis selected based on colour cuts that favour the detection ofsystems containing cooler WDs and/or earlier type compan-ions. This may be a consequence of the fact that it is moredifficult to derive reliable WD parameters for this type ofsystems. That is, cooler WDs are systematically fainter andhence their spectra are of lower SN ratio, which translatesinto larger parameter uncertainties. Hence the objects con-taining the coolest WDs either cannot be fitted because theyare subject to too low SN spectra or the fitted values do notsurvive our quality cuts and are excluded from the analysis.This, unfortunately, implies that our sample of LAMOSTDR5 DA/M binaries with available stellar parameters is stillbiased against systems containing cool WDs.

We also run KS and χ2 tests to the cumulative dis-tributions resulting from the entire LAMOST DR5 DA/Msample and the LAMOST DR1 catalogue we presented inRen et al. (2014). The distributions are shown in the rightpanels of Figure 8 and in Table 2 we provide the probabili-ties. Again, we find no strong indications for the DR1 andDR5 LAMOST populations to be statistically different.

0 1000 2000 3000 4000 5000

time span (d)

0

100

200

300

400

maxim

um

RV

shift

(km

/s) NaI doublet

0 1000 2000 3000 4000 5000

time span (d)

Hα

Figure 9. Left panel: The maximum Na I doublet RV shift mea-sured vs. time span between the observed spectra (in days) forthe 76 new PCEBs identified in this work. Right panel: The samebut for Hα emission. The numerical values of the RV shifts areprovided in Table 4.

7 IDENTIFICATION OF CLOSE WDMSBINARIES

In this section we aim at detecting PCEBs among theWDMS binaries identified in this work. To that end we re-quire measuring the radial velocities (RVs) of at least onestellar component. WD RVs can be measured fitting theBalmer lines (Breedt et al. 2017) and/or employing cross-correlation techniques (Anguiano et al. 2017). However, weexpect the WD RV shifts to be smaller than those measuredfrom their M dwarf companions, since the WDs are gener-ally closer to the center of mass in this type of binaries. Wehence measure the Na Iλλ 8183.27, 8194.81 absorption dou-blet and the Hα emission RVs arising from the M dwarfcompanions from all the LAMOST WDMS binary spectraidentified in this work. We identify a PCEB when we detectmore than 3σ RV variation. Conversely, if no RV variation isdetected from spectra taken separated by at least one night,the system is considered as a likely wide binary candidatethat presumably evolved avoiding episodes of mass transfer.We fit the Na Iλλ 8183.27, 8194.81 absorption doublet with a

Downloaded from https://academic.oup.com/mnras/advance-article-abstract/doi/10.1093/mnras/sty805/4956533by gueston 12 April 2018

10 J.-J. Ren et al.

Table 3. Radial velocities measured from the Na Iλλ 8183.27, 8194.81 absorption doublet and the Hα emission for the LAMOST DR5WDMS binaries. The Heliocentric Julian Dates (HJD) are also indicated. The last column lists the telescope used for obtaining thespectra (either LAMOST or SDSS). We use ‘−’ to indicate that no radial velocity is available. The complete table can be found in theelectronic version of the paper.

Jname HJD RV(Na I) Err RV(Hα) Err Telescope(days) (km s−1) (km s−1) (km s−1) (km s−1)

J000448.23+343627.4 2456948.07023 −21.59 45.97 −96.14 17.32 LAMOSTJ000448.23+343627.4 2456948.09106 −45.61 20.86 −93.60 18.83 LAMOSTJ000448.23+343627.4 2456948.09341 −43.07 15.78 −62.27 24.84 LAMOSTJ000448.23+343627.4 2456948.11661 −37.64 20.67 −76.21 19.09 LAMOSTJ000612.32+340358.3 2456948.07026 −4.35 12.62 −12.92 10.50 LAMOSTJ000612.32+340358.3 2456948.09109 39.19 12.72 14.71 10.45 LAMOSTJ000612.32+340358.3 2456948.09344 30.75 11.87 9.55 10.28 LAMOSTJ000612.32+340358.3 2456948.11664 64.64 12.63 26.05 10.43 LAMOSTJ000729.32+023124.5 2455892.99443 14.98 14.38 − − LAMOST

Table 4. The 76 new PCEBs found in this work. The Flag column indicates whether the PCEBs are detected based on Na I and/orHα emission RV measurements (only from Na I: 10; only from Hα; 01 ; from both Na I and Hα: 11). The maximum RV shifts are alsoindicated.

Jname Flag RV shift RV shift Jname Flag RV shift RV shift Jname Flag RV shift RV shift(km/s) (km/s) (km/s) (km/s) (km/s) (km/s)Na I Hα Na I Hα Na I Hα

J000612.32+340358.3 10 68.99 − J093507.99+270049.2 11 248.88 182.41 J122514.92+292523.1 11 113.00 87.11J004751.47+340212.7 01 − 70.77 J093809.28+143036.9 11 100.51 113.59 J123339.40+135943.2 11 133.43 106.79J011739.90+340209.4 11 177.57 158.24 J094101.91+510719.8 01 − 59.05 J123642.53+580230.6 11 86.16 69.09J011847.16+050531.2 11 155.83 100.75 J095641.79+013130.7 01 − 87.01 J124831.79+435318.3 11 78.02 121.40J024924.76+071344.3 01 − 118.56 J101356.31+272410.8 11 600.06 652.14 J124830.76+540803.6 11 83.12 219.67J025301.60−013006.8 01 − 72.39 J101616.82+310506.5 11 114.66 66.68 J132001.24+112805.3 10 79.95 −

J030944.99+285507.5 10 58.73 − J103125.93+020751.4 11 76.84 61.33 J132341.89+541636.5 10 137.35 −

J032750.32+222618.9 10 109.05 − J104318.65+305012.9 10 146.37 − J134051.00+321015.9 01 − 127.69J050825.81+252050.8 11 102.87 73.50 J104534.46+315740.6 01 − 80.16 J135825.68+171204.1 11 377.78 235.72J054119.24−051838.4 11 154.71 167.96 J105352.90+340923.5 01 − 254.71 J140115.68+214157.1 10 49.64 −

J060040.31+312118.6 10 45.48 − J105657.35+330416.2 11 102.26 68.60 J142917.86+285103.0 10 90.38 −

J063840.55+130253.5 01 − 101.87 J110827.40+303031.3 01 − 73.91 J144307.83+340523.5 01 − 94.57

J073128.30+264353.6 11 56.08 161.49 J111614.48+535720.9 01 − 74.54 J145248.79+234807.6 01 − 54.60J081126.70+053912.9 01 − 293.08 J112007.64+250221.3 11 193.70 204.91 J145430.02+203902.5 11 120.18 148.83J081654.35+354230.6 10 146.69 − J112124.82+273046.2 11 256.30 252.29 J151042.63+273509.9 10 90.99 −

J082214.50+433343.6 10 98.77 − J113102.81+522645.3 01 − 74.78 J153934.63+092221.5 11 175.68 118.65J082656.51+480545.7 11 148.67 145.12 J113910.49+315013.1 10 121.59 − J153914.69+263710.8 01 − 290.97J082845.07+133551.6 11 227.51 210.96 J114732.48+591735.1 10 60.98 − J154101.07+294829.4 11 168.15 185.74J083056.11+315941.9 01 − 69.67 J114732.25+593921.9 10 106.47 − J160645.02+284725.9 11 88.25 100.14J084418.12+340458.9 11 122.49 105.03 J115057.60+282637.8 11 235.92 160.12 J165354.16+322937.6 01 − 75.64J085024.05+054757.8 10 37.21 − J115026.28+302359.4 01 − 231.17 J213136.72+002947.6 10 74.41 −

J085025.75+344053.2 10 98.86 − J115113.14−011111.1 01 − 168.85 J224522.49+063817.2 01 − 88.17J085414.26+211148.1 11 233.93 149.53 J115510.86+271324.3 01 − 95.53 J224805.92+144328.3 01 − 66.04J085634.83+373913.4 11 180.94 202.61 J121301.82+282310.0 11 389.86 388.75 J234106.82+083550.3 01 − 81.32J090210.95+252913.3 01 − 114.09 J122037.01+492334.0 10 76.86 − J235408.04+291623.1 01 − 54.65J092712.02+284629.2 01 − 96.31

second-order polynomial plus a double-Gaussian line profileof fixed separation, while the Hα emission line (if present inthe spectrum) is fitted with a second-order polynomial plus asingle-Gaussian line profile (Rebassa-Mansergas et al. 2008;Ren et al. 2013). With the aim of increasing the SN ratio,the final released LAMOST DR5 spectra are the result ofcombining several different sub-exposure spectra (hereaftersub-spectra). Hence, we measure the RVs from all availablesub-spectra as well as from the final combined spectra.

We cross-matched our LAMOST DR5 catalogue withthe current newest spectroscopic data base of SDSS, i.e.DR14. This resulted in 685 additional spectra for 465 ofour targets. Combining LAMOST DR5 and SDSS DR14 wethus gathered a total of 1835 spectra (1150 from LAMOST,

685 from SDSS) for measuring RVs. We derived at leastone Na Iλλ 8183.27, 8194.81 absorption doublet and one Hα

emission RV of an accuracy better than 20 kms−1 for 685and 477 WDMS binaries in our catalogue, respectively. For607 and 408 WDMS binaries we managed to measure atleast two Na I and Hα RVs, respectively. The RVs are pro-vided in Table 3 (for completeness, we also provide the RVswe measured with accuracies worse than 20 kms−1).

We identified 89 and 103 objects displaying more than3σ Na I and Hα RV variation, respectively. If we only takeinto account systems with RVs taken on different nights (337objects for Na I, 229 objects for Hα), we obtain a PCEBfraction of ∼26 per cent based on the Na I RV measure-ments, or ∼ 45 per cent based on the Hα RV measure-

Downloaded from https://academic.oup.com/mnras/advance-article-abstract/doi/10.1093/mnras/sty805/4956533by gueston 12 April 2018

WDMS from LAMOST DR5 11

Table 5. LAMOST DR5 WDMS binaries that display light curve variations as revealed by the Catalina photometric surveys. We providenotes on individual systems, classification and orbital periods obtained by calculating periodograms from the photometric data.

Object Notes Classification Orbital Period

J000729.32+023124.5 Some faint points Candidate new eclipser ?J001752.63+332424.9 Reflection effect? Candidate non-eclipser 6.4 hours?J003033.11+070657.5 Seems to display variation Candidate non-eclipser 1.2 hours? or doubleJ004232.56+415403.0 Reflection effect? Non-eclipser 0.10312 daysJ004751.47+340212.7 Large variations Candidate non-eclipser ?J005130.56+314555.9 Reflection effect? Candidate non-eclipser 1.7 hours?J010010.53+003739.1 Reflection effect? Candidate non-eclipser 1.7 hours?J011547.58+005350.0 Not many points Candidate non-eclipser 1.2 hours?J012549.89+330940.6 Variations Non-eclipser 0.24244 days, or doubleJ012550.48+280756.0 Some faint points Candidate new eclipser ?J013157.96+084948.2 Reflection effect? Candidate non-eclipser 2.3 hours?J024924.77+071344.3 Variations Non-eclipser 0.17334 days, or doubleJ030308.35+005444.1 Clear eclipser Known eclipser [1] 0.13444 daysJ042955.26+344734.3 Clear eclipser New eclipser 0.30716 daysJ064959.81+424110.2 Reflection effect? Non-eclipser 0.24496 daysJ082145.27+455923.4 Clear eclipser Known eclipser [2] 0.50909 daysJ084028.85+501238.2 Variation Candidate non-eclipser 1.2 hours? or doubleJ085414.28+211148.2 Reflection effect? Non-eclipser 0.10215 daysJ085835.56+281356.3 large variations LARP? ?J090812.04+060421.2 Clear eclipser Known eclipser [3] 0.14944 daysJ091216.37+234442.5 Reflection effect? Non-eclipser 0.26356 daysJ092712.02+284629.2 Variation Candidate non-eclipser 3.1 hours? or doubleJ092741.73+332959.1 Clear eclipser Known eclipser [2] 2.30822 daysJ093207.63+334805.9 Some faint points Candidate new eclipser 1.7 days?J093507.99+270049.2 Clear eclipser Known eclipser [4] 0.20103 daysJ093947.95+325807.3 Clear eclipser Known eclipser [3] 0.33098 daysJ094913.36+032254.5 Reflection effect? Candidate non-eclipser 3.1 hours?J095719.24+234240.7 Clear eclipser Known eclipser [3] 0.15087 daysJ095737.59+300136.5 Clear eclipser Known eclipser [2] 1.92613 daysJ101307.79+245713.1 Clear eclipser Known eclipser [5] 0.12904 daysJ104012.99+252559.9 Reflection effect? Candidate non-eclipser 3.7 hours?J110827.40+303031.3 Variation Non-eclipser 0.81820 days, or doubleJ112007.64+250221.3 Some faint points Candidate new eclipser ?J112738.71+281532.7 Several faint points Candidate new eclipser ?J113102.81+522645.3 Variation Candidate non-eclipser 2 hours? or double

J114224.71−022610.0 Variation Non-eclipser 0.56112 days, or doubleJ114509.77+381329.2 Variation Non-eclipser 0.19004 days, or doubleJ114853.34+555217.0 Active star? Flares? Active M dwarf? ?J120020.81+363557.3 Some faint points Candidate new eclipser ?J121010.13+334722.9 Clear eclipser Known eclipser [6] 0.12449 daysJ121258.25−012309.9 Clear eclipser Known eclipser [7] 0.33587 daysJ122630.86+303852.5 Reflection effect? Non-eclipser 0.25869 daysJ123214.38+351324.8 Variation Candidate non-eclipser 2.5 hours?J132925.21+123025.4 Clear eclipser Known eclipser [2] 0.08097 daysJ134234.78+304849.2 Several faint points Candidate new eclipser 1.66 days?J135825.68+171204.1 Variation Non-eclipser 0.16046 days, or doubleJ141811.97+204150.8 Reflection effect? Candidate non-eclipser 2.6 hours?J143547.87+373338.5 Clear eclipser Known eclipser [8] 0.12563 daysJ143900.62+560219.0 Variation Non-eclipser 0.34268 days, or doubleJ144307.83+340523.5 Shows big variations Unclear ?J144846.85+071304.3 Some faint points, real? Candidate new eclipser ?J151426.90+285720.4 Flares? CV? Unclear ?J152804.92+331012.1 Clear eclipser New eclipser 0.21155 daysJ154846.00+405728.7 Clear eclipser Known eclipser [8] 0.18552 daysJ212309.40+040929.5 Many faint points Candidate new eclipser ?J212531.92−010745.9 Reflection effect Non-eclipser 0.28982 daysJ233900.38+115707.2 Variation Non-eclipser 0.12286 days, or double

[1] Parsons et al. (2013b); [2] Parsons et al. (2013a); [3] Drake et al. (2010); [4] Drake et al. (2014); [5] Parsons et al. (2015a); [6]Pyrzas et al. (2012); [7] Parsons et al. (2012b); [8] Pyrzas et al. (2009)

Downloaded from https://academic.oup.com/mnras/advance-article-abstract/doi/10.1093/mnras/sty805/4956533by gueston 12 April 2018

12 J.-J. Ren et al.

Figure 10. Left panels: CSS light curves folded over the bestorbital period determinations (see right panels) of the two neweclipsing DA/M binaries identified in this work. Top right panel:the results of a standard Lomb-Scargle periodogram applied tothe CSS photometry of J042955.26+344734.3, peaking at 3.25556cycles/day (corresponding to a period of 0.30716 d). Bottom rightpanel: displayed are the results obtained applying a box-searchmethod (see Section 8 for details) to the CSS photometry ofJ152804.92+331012.1, resulting in an orbital period of 0.21155days.

ments. The PCEB fraction derived from the Na I RV anal-ysis agrees well with those obtained in previous works fromthe SDSS WDMS sample (e.g. Nebot Gomez-Moran et al.2011; Rebassa-Mansergas et al. 2016a). It has to be notedthat the higher PCEB fraction derived by analysing the Hα

RVs needs to be taken with caution. This is because Hα RVsare less robust for detecting RV variations due to magneticactivity effects (Rebassa-Mansergas et al. 2008).

The total number of unique PCEBs we have identifiedthat results from excluding duplicated targets from both theNa I and Hα PCEB lists is 128. Among these, four were pre-viously identified by Ren et al. (2014) within the LAMOSTDR1 catalogue and 50 by Rebassa-Mansergas et al. (2016a)within the SDSS DR12 sample (we note that two were bothdetected within the LAMOST DR1 and SDSS DR12 cata-logues). Thus, we are left with 76 new PCEB identificationsin this work. In table 4 we provide the object names of these76 PCEBs and in Figure 9 we show their maximum radialvelocity shift vs. time span between the observed spectra. Itis worth mentioning that in ∼1/2 of the cases the time spanis over 1000 days. This is due to the fact that we combineLAMOST and SDSS spectra to detect close binaries, thusresulting in time baselines as long as ∼15 years.

8 ECLIPSING WDMS SYSTEMS

Following the approach presented in Parsons et al. (2013a)and Parsons et al. (2015a), we cross-matched our LAMOSTDR5 WDMS binary catalogue with the photometric datafrom the Catalina Sky Survey (CSS) and Catalina Real TimeTransient Survey (CRTS Drake et al. 2009). This alloweddetecting WDMS binaries displaying light curve variations.During this process the raw CSS data were re-reduced by us

to identify deeply eclipsing systems and to remove contam-inated exposures.

Among the 876 unique WDMS binaries that form theDR5 LAMOST catalogue, 687 were observed by the Catalinasurveys, of which 630 did not display light curve variations.The remaining 57 objects include 16 eclipsing systems, twoof which are new (see the light curves in Figure 10), 9 neweclipsing candidates, 14 objects displaying reflection effectsor irradiation effects, 14 candidates for displaying reflectionor irradiation effects, one candidate low-accretion rate polar,one candidate active M dwarf and two systems displayinglight curve variations which we are not able to classify. Weprovide additional notes and our classification for these 57objects in Table 5.

We calculated Scargle (1982) periodograms from theCatalina photometric data with the aim of measuring theorbital periods of the 57 objects displaying light curve vari-ations. This is a suitable method when analysing light curvesdisplaying out-of-eclipse variations due to reflection or ellip-soidal modulation effects (see an example in the top rightpanel of Figure 10). In the absence of such variations, i.e.only the eclipses are sampled by the CSS data, we employedthe box fitting method outlined by Parsons et al. (2013a).We first identify the in-eclipse points from the raw lightcurve and fold the data over a given orbital period. Fromthe folded data we identify the first and last in-eclipse pointsand count the number of out-of-eclipse points in between.This procedure is repeated over a large range of adopted or-bital periods, being the correct determination the value thatresults in zero out-of-eclipse points. The number of out-of-eclipse points as a function of the adopted orbital period forour newly identified eclipsing system J152804.92+331012.1is shown in the bottom right panel of Figure 10.

For 30 systems we were able to determine precise val-ues of the orbital periods. In additional 15 cases, we couldonly derive estimates of the orbital periods due to insuffi-cient information provided by the periodograms. We notethat since, a priory, we do not know whether the light curvevariations are due to reflection of irradiation effects for ournon-eclipsing systems, there is a possibility that the orbitalperiods are double the measured ones. We include our mea-sured values of the orbital periods in Table 5. We note that,for the two new eclipsing systems found here, we were notable to derive accurate RVs due to the low SN ratio of theirspectra, thus they are not included in our PCEB list.

9 SUMMARY AND CONCLUSIONS

The catalogue of WDMS binaries from LAMOST DR5 con-tains 876 objects and it is ∼8 times larger than our previousDR1 sample. 357 of these systems (∼40 per cent of the cat-alogue) are new identifications that have not been publishedbefore. Moreover, 339 were observed as part of a dedicatedLAMOST survey for obtaining spectra of WDMS binariesphotometrically selected within SDSS that are expected tocontain cool WDs and/or early type M dwarf companions.

We determined the stellar parameters (white dwarf ef-fective temperatures, surface gravities and masses, and Mdwarf spectral types) of our systems following a decompo-sition/fitting routine, and we used the corresponding pa-rameter distributions to analyse the intrinsic properties of

Downloaded from https://academic.oup.com/mnras/advance-article-abstract/doi/10.1093/mnras/sty805/4956533by gueston 12 April 2018

WDMS from LAMOST DR5 13

the LAMOST DR5 sample. We found that the populationof cool WDs remains under-represented. This is most likelydue to the fact that cool (Teff.10 000 K) WDs are systemat-ically fainter and hence associated to lower SN ratio spectra.This increases considerably the probability for our decom-position/fitting method to determine WD parameters asso-ciated to large uncertainties and, as a consequence, theseobjects are not taken into account in our analysis.

We measured the Na Iλλ 8183.27, 8194.81 absorptiondoublet and the Hα emission radial velocities of each ob-ject in our catalogue from their LAMOST DR5 as well asSDSS DR14 spectra (when available). We detected 128 sys-tems (76 of which are new identifications) displaying morethan 3σ radial velocity variations and hence classify theseobjects as PCEBs. The close binary fraction we derived is∼26 per cent, in agreement with previous studies.

By cross-matching our catalogue with the Catalina Sur-veys we found 57 systems displayed light curve variations.Among these we identified 16 eclipsing systems, two of whichare new, and nine additional eclipsing WDMS binary can-didates. By analysing the periodograms calculated from thephotometric data we were able to determine the orbital pe-riods of 30 objects and estimate the orbital periods of 15additional systems.

ACKNOWLEDGEMENTS

We thank the anonymous referee for the helpful commentsand suggestions. This work is supported by Joint Funds ofthe National Natural Science Foundation of China (GrantNo. U1531244) and the National Key Basic Research Pro-gram of China 2014CB845700. JJR acknowledges supportfrom the Young Researcher Grant of National Astronomi-cal Observatories, Chinese Academy of Sciences. ARM ac-knowledges support from the MINECO under the Ramony Cajal programme (RYC-2016-20254) and the AYA2017-86274-P grant, and the AGAUR grant SGR-661/2017.

This work has made use of data products from theGuoshoujing Telescope (the Large Sky Area Multi-ObjectFibre Spectroscopic Telescope, LAMOST). LAMOST is aNational Major Scientific Project built by the ChineseAcademy of Sciences. Funding for the project has been pro-vided by the National Development and Reform Commis-sion. LAMOST is operated and managed by the NationalAstronomical Observatories, Chinese Academy of Sciences.

REFERENCES

Anguiano B., Rebassa-Mansergas A., Garcıa-Berro E., Torres S.,Freeman K. C., Zwitter T., 2017, MNRAS, 469, 2102

Bours M. C. P., Marsh T. R., Parsons S. G., et al., 2016, MNRAS,460, 3873

Breedt E., et al., 2017, MNRAS, 468, 2910Camacho J., Torres S., Garcıa-Berro E., Zorotovic M., Schreiber

M. R., Rebassa-Mansergas A., Nebot Gomez-Moran A., Gan-sicke B. T., 2014, A&A, 566, A86

Chui C. K., 1992, Wavelets: A tutorial in theory and applicationsCojocaru R., Rebassa-Mansergas A., Torres S., Garcıa-Berro E.,

2017, MNRAS, 470, 1442Cui X.-Q., et al., 2012, Research in Astronomy and Astrophysics,

12, 1197Davis P. J., Kolb U., Willems B., 2010, MNRAS, 403, 179

De Marco O., Passy J.-C., Moe M., Herwig F., Mac Low M.-M.,

Paxton B., 2011, MNRAS, 411, 2277

Deng L.-C., et al., 2012, Research in Astronomy and Astrophysics,12, 735

Drake A. J., et al., 2009, ApJ, 696, 870

Drake A. J., et al., 2010, preprint, (arXiv:1009.3048)

Drake A. J., et al., 2014, ApJS, 213, 9

Farihi J., Hoard D. W., Wachter S., et al., 2010, ApJS, 190, 275

Gansicke B. T., et al., 2009, MNRAS, 397, 2170

Gentile Fusillo N. P., et al., 2015, MNRAS, 452, 765

Han Z., Podsiadlowski P., Maxted P. F. L., Marsh T. R., IvanovaN., 2002, MNRAS, 336, 449

Han Z., Podsiadlowski P., Maxted P. F. L., Marsh T. R., 2003,MNRAS, 341, 669

Han Z., Podsiadlowski P., 2004, MNRAS, 350, 1301

Heller R., Homeier D., Dreizler S., Østensen R., 2009, A&A,496, 191

Heber U., 2016, PASP, 128, 082001

Iben Jr. I., Livio M., 1993, PASP, 105, 1373

Kepler S. O., et al., 2015, MNRAS, 446, 4078

Kilic M., Brown W. R., Gianninas A., Curd B., Bell K. J., AllendePrieto C., 2017, MNRAS, 471, 4218

Koester D., 2010, Mem. Soc. Astron. Italiana, 81, 921

Koester D., Kepler S. O., Kleinman S. J., Nitta A.,2009, in Journal of Physics Conference Series. p. 012006(arXiv:0812.0491), doi:10.1088/1742-6596/172/1/012006

Langer N., Deutschmann A., Wellstein S., Hoflich P., 2000, A&A,362, 1046

Li L., Zhang F., Han Q., Kong X., Gong X., 2014, MNRAS,445, 1331

Liu C., Li L., Zhang F., Zhang Y., Jiang D., Liu J., 2012, MNRAS,424, 1841

Liu X.-W., et al., 2014, in Feltzing S., Zhao G., Walton N. A.,Whitelock P., eds, IAU Symposium Vol. 298, Setting thescene for Gaia and LAMOST. pp 310–321 (arXiv:1306.5376),doi:10.1017/S1743921313006510

Luo A.-L., et al., 2012, Research in Astronomy and Astrophysics,12, 1243

Luo A.-L., et al., 2015, Research in Astronomy and Astrophysics,15, 1095

Marsh T. R., et al., 2014, MNRAS, 437, 475

Miszalski B., Acker A., Moffat A. F. J., et al., 2009, A&A,496, 813

Nebot Gomez-Moran A., et al., 2009, A&A, 495, 561

Nebot Gomez-Moran A., et al., 2011, A&A, 536, A43

Pala A. F., et al., 2017, MNRAS, 466, 2855

Parsons S. G., et al., 2012a, MNRAS, 419, 304

Parsons S. G., et al., 2012b, MNRAS, 420, 3281

Parsons S. G., et al., 2013a, MNRAS, 429, 256

Parsons S. G., Marsh T. R., Gansicke B. T., Schreiber M. R.,Bours M. C. P., Dhillon V. S., Littlefair S. P., 2013b, MNRAS,436, 241

Parsons S. G., et al., 2014, MNRAS, 438, L91

Parsons S. G., et al., 2015a, MNRAS, 449, 2194

Parsons S. G., et al., 2015b, MNRAS, 452, 1754

Parsons S. G., et al., 2017, MNRAS, 470, 4473

Passy J.-C., et al., 2012, ApJ, 744, 52

Pyrzas S., et al., 2009, MNRAS, 394, 978

Pyrzas S., et al., 2012, MNRAS, 419, 817

Pyrzas S., et al., 2015, MNRAS, 447, 691

Rebassa-Mansergas A., Gansicke B. T., Rodrıguez-Gil P.,Schreiber M. R., Koester D., 2007, MNRAS, 382, 1377

Rebassa-Mansergas A., et al., 2008, MNRAS, 390, 1635

Rebassa-Mansergas A., Gansicke B. T., Schreiber M. R., KoesterD., Rodrıguez-Gil P., 2010, MNRAS, 402, 620

Rebassa-Mansergas A., Nebot Gomez-Moran A., Schreiber M. R.,Girven J., Gansicke B. T., 2011, MNRAS, 413, 1121

Downloaded from https://academic.oup.com/mnras/advance-article-abstract/doi/10.1093/mnras/sty805/4956533by gueston 12 April 2018

14 J.-J. Ren et al.

Rebassa-Mansergas A., Nebot Gomez-Moran A., Schreiber M. R.,

Gansicke B. T., Schwope A., Gallardo J., Koester D., 2012a,MNRAS, 419, 806

Rebassa-Mansergas A., et al., 2012b, MNRAS, 423, 320Rebassa-Mansergas A., Schreiber M. R., Gansicke B. T., 2013a,

MNRAS, 429, 3570Rebassa-Mansergas A., Agurto-Gangas C., Schreiber M. R., Gan-

sicke B. T., Koester D., 2013b, MNRAS, 433, 3398Rebassa-Mansergas A., Ren J. J., Parsons S. G., Gansicke B. T.,

Schreiber M. R., Garcıa-Berro E., Liu X.-W., Koester D.,2016a, MNRAS, 458, 3808

Rebassa-Mansergas A., et al., 2016b, MNRAS, 463, 1137Rebassa-Mansergas A., Parsons S. G., Garcıa-Berro E., Gan-

sicke B. T., Schreiber M. R., Rybicka M., Koester D., 2017,MNRAS, 466, 1575

Ren J., Luo A., Li Y., Wei P., Zhao J., Zhao Y., Song Y., ZhaoG., 2013, AJ, 146, 82

Ren J. J., et al., 2014, A&A, 570, A107Renedo I., Althaus L. G., Miller Bertolami M. M., Romero A. D.,

Corsico A. H., Rohrmann R. D., Garcıa-Berro E., 2010, ApJ,717, 183

Richards G. T., et al., 2002, AJ, 123, 2945Ricker P. M., Taam R. E., 2012, ApJ, 746, 74Scargle J. D., 1982, ApJ, 263, 835Schleicher D. R. G., Dreizler S., Volschow M., Banerjee R., Hes-

sman F. V., 2015, Astronomische Nachrichten, 336, 458Schreiber M. R., Gansicke B. T., Southworth J., Schwope A. D.,

Koester D., 2008, A&A, 484, 441Schreiber M. R., et al., 2010, A&A, 513, L7Silvestri N. M., et al., 2006, AJ, 131, 1674Skinner J. N., Morgan D. P., West A. A., Lepine S., Thorstensen

J. R., 2017, AJ, 154, 118Smolcic V., et al., 2004, ApJ, 615, L141Stoughton C., et al., 2002, AJ, 123, 485Su D.-Q., Cui X.-Q., 2004, Chinese J. Astron. Astrophys., 4, 1Toonen S., Nelemans G., 2013, A&A, 557, A87Tremblay P.-E., Ludwig H.-G., Steffen M., Bergeron P., Freytag

B., 2011, A&A, 531, L19Tremblay P.-E., Ludwig H.-G., Steffen M., Freytag B., 2013,

A&A, 559, A104Veras D., Xu S., Rebassa-Mansergas A., 2018, MNRAS, 473, 2871Wang B., Han Z., 2012, New Astron. Rev., 56, 122Wang B., Li X.-D., Han Z.-W., 2010, MNRAS, 401, 2729Wang S.-G., Su D.-Q., Chu Y.-Q., Cui X., Wang Y.-N., 1996,

Appl. Opt., 35, 5155Webbink R. F., 2008, in Milone E. F., Leahy D. A., Hobill D. W.,

eds, Astrophysics and Space Science Library Vol. 352, Astro-

physics and Space Science Library. p. 233 (arXiv:0704.0280),doi:10.1007/978-1-4020-6544-6 13

Willems B., Kolb U., 2004, A&A, 419, 1057Xiang M.-S., et al., 2017, MNRAS, 467, 1890Yanny B., et al., 2009, AJ, 137, 4377York D. G., et al., 2000, AJ, 120, 1579Yuan H.-B., et al., 2015, MNRAS, 448, 855Zhao G., Zhao Y.-H., Chu Y.-Q., Jing Y.-P., Deng L.-C., 2012,

Research in Astronomy and Astrophysics, 12, 723Zorotovic M., Schreiber M. R., 2013, A&A, 549, A95Zorotovic M., Schreiber M. R., Gansicke B. T., Nebot Gomez-

Moran A., 2010, A&A, 520, A86Zorotovic M., Schreiber M. R., Garcıa-Berro E., Camacho J., Tor-

res S., Rebassa-Mansergas A., Gansicke B. T., 2014, A&A,568, A68

Zorotovic M., et al., 2016, MNRAS, 457, 3867de Kool M., 1992, A&A, 261, 188

This paper has been typeset from a TEX/LATEX file prepared bythe author.

Downloaded from https://academic.oup.com/mnras/advance-article-abstract/doi/10.1093/mnras/sty805/4956533by gueston 12 April 2018